Alzheimer's disease (AD) is the most common cause of dementia and is the third leading cause of death in older adults in the US. Over 4.7 million Americans aged 65 and older live with AD and this number is expected to climb to nearly 14 million by 2050. However, fewer than 50% of individuals with AD have been diagnosed. and patients who have the condition are unlikely to receive a full diagnostic workup to identify the cause(s) of their impairment. Detection of functional markers is an essential first step toward prevention and early diagnosis of MCI, Alzheimer's Disease (AD) and related dementias. Recently, several highly regarded clinical aging studies have demonstrated that subtle changes in gait are early, sensitive and specific, noninvasive risk markers for both cognitive decline and fall risk. The majority of these clinical investigations have used an instrumented mat system (e.g. the GaitRite system) that measures spatio-temporal gait parameters from footfalls. Although the literature is consistent about the potential for spatio-temporal footfall-based metrics as a predictor of MCI and dementia, the technology industry has not delivered an accurate yet affordable solution appropriate for widespread use in a primary care setting. A new and novel approach is required to transfer the significant clinical research findings to clinical practice. The goal of this SBIR is to create an accurate, low cost, simple-to-use primary care clinical screening tool for MCI and dementia and a risk assessment and stratification tool for older adults with normal cognition. This will be accomplished by commercializing recent research in novel software-based deep learning marker-less motion capture and advanced kinematic analysis methods. A secure mobile application and cloud-based big data analytics platform delivers this Software-as-a Service to providers with the low cost and ease-of use required to accelerate adoption. This new approach applies deep learning technology to measure footfall-based gait parameters. It can replicate the level of precision and accuracy of expensive (>$35,000) and space-consuming electronic mats used in previous research studies. In addition, our work will significantly advance the research field by simultaneously measuring, from the same video stream, accurate 3D joint angles, which have recently been shown to be even more specific markers of neurodegeneration compared to footfall parameters. Long Term Goal: All patients will routinely have equal access to advanced gait analytics in primary care practices. By providing a consistent methodology, including integration in the Medicare Annual Wellness Visit, very large population data will be collected and analyzed to provide new insights into early stages of dementia, discover new functional markers from 3D kinematics, improve diagnostic assessments, and identify new preventive strategies for cognitive decline and risk of falls.